OpenAI’s CEO explained that he was “terrified” that ChatGPT could be used for massive disinformation and offensive cyberattacks.
Sam Altman recently spoke with ABC News about the release of the latest version of the company’s chatbot and AI language model, GPT-4.
And while the chatbot has raised fears that artificial intelligence will dominate the world, Altman believes that humans are the technology’s biggest threat.
“There are going to be other people who don’t put some of the safety boundaries that we do. I think society has limited time to understand how to respond, how to fix this,” he told ABC News. it’s how to deal with it.”
OpenAI released the GPT-4 model last week and described it as more powerful than its predecessor – so it could be “damaging”.
The bot accepts input in image format as well as text, but still gives its responses as text, meaning it can provide detailed descriptions of images.
The company warns that the model tends to “hallucinate” false facts and may be induced to vomit false or harmful content.
“The thing I try to warn people about most is what we call the ‘hallucination problem,’” Altman said. “The model has the confidence to explain things as if they were purely fictional.”
In the interview, Altman admitted that the GPT-4 is “not perfect,” but it can write computer code in most programming languages and pass the standard tape test at the 90th percentile.
However, Altman sees GPT-4 as an evil in the hands of those who use its power for evil.
“I am deeply concerned that these models could be used for large-scale disinformation. Now that you’re better at writing computer code, they can be used in offensive cyber attacks,” Altman told ABC News.
He also mentioned another fear that is spreading around the world – that artificial intelligence will take over human jobs.
Altman doesn’t rule out that technology will replace some jobs, and he believes it could happen faster than we think.
“It has been proven, over several generations, that humanity can adapt amazingly to major technological changes,” Altman said. “
GPT-4 is now available to ChatGPT subscribers: free users will still encounter the old GPT-3.5.
Speaking at SXSW this month, OpenAI co-founder Greg Brockman said fears that AI tools will take people’s jobs are overblown, and that AI will free people to focus on what matters. task.
“The most important thing is that these are high-level skills – reasoning, knowing when to go into details. I think the real story here is expanding what people can do.”
Source: Daily Mail
Source: Arabic RT
